SunflowersLw AI & On-Device Developer ToolsAI 与设备端开发者工具.
Tools that live where developers work — the IDE, the terminal, the device.住在开发者工作之处的工具——IDE、终端、设备。
Developer tools with an AI and on-device edge: an open-source MCP sidecar that turns your AI code changes into quizzes and remembers your debugging context, plus a native document renderer with an on-device git client. Built to plug into real workflows rather than replace them.兼具 AI 与设备端优势的开发者工具：一个开源的 MCP 边车，把你的 AI 代码改动变成测验，并记住你的调试上下文；外加一个带设备端 git 客户端的原生文档渲染器。它们被打造来接入真实的工作流，而不是取代它们。
covate
Dev Tool · MCP开发者工具 · MCPOpen-source MCP learning sidecar and persistent debugging memory for AI assistants.面向 AI 助手的开源 MCP 学习边车与持久化调试记忆。
- Quizzes you on your own AI-assisted changes when you ask在你提出时，就你自己的 AI 辅助代码改动出题考你
- Persistent debugging memory across coding sessions跨会话持久化记录调试思路与排错路径，避免重复踩坑
- MCP sidecar that plugs directly into Claude Code, Codex, and modern AI tools无缝接入 Claude Code、Codex 及标准 MCP 协议编程宿主
FoldRender
Android · Foldable安卓 · 折叠屏Native document renderer and on-device git client for mobile and foldable screens.面向移动端与折叠屏的原生文档渲染器与设备端 Git 客户端。
- Native renderer for the repo reader (not a WebView)仓库阅读器为原生渲染（不是 WebView）
- Unfolding foldable devices triggers pure relayout with zero reflow jank针对大折叠屏展开做纯重排优化，流畅无重绘卡顿
- On-device git: clone and pull (diff, commit and push are in the build, not yet verified end to end)端上 Git：克隆与拉取（差异、提交、推送已在包内，尚未完整验证）
FAQ常见问题
What are useful AI tools for developers?对开发者有用的 AI 工具有哪些？
covate is an open-core learning sidecar: a free MCP that quizzes you on your own AI-assisted code changes and keeps persistent debugging memory across sessions. FoldRender is a native document renderer with on-device clone and pull, built for foldable Android.covate 是一个开源核心的学习边车：一个免费的 MCP，就你自己借助 AI 完成的代码改动向你提问，并在多次会话之间保留持久的调试记忆。FoldRender 是一个原生文档渲染器，带设备端的克隆和拉取，为折叠屏 Android 打造。
What is an MCP sidecar?什么是 MCP 边车？
MCP (Model Context Protocol) lets AI coding assistants call external tools. covate ships as an MCP that plugs into your assistant to reinforce learning from the code changes the AI makes with you.MCP（Model Context Protocol）让 AI 编程助手能够调用外部工具。covate 以 MCP 的形式运行，从而接入你的助手，并就 AI 与你一起做出的代码改动来巩固学习。